False data injection attack detection of power grid based on IWOA-BiLSTM

False Data Injection Attacks (FDIA) disrupt or mislead the decision-making of the command center by injecting covert false data into the monitoring instrument to bypass the bad data detection mechanism. In this paper, a false data injection attack testing method based on IWOA-BiLSTM is proposed. The proposed method uses a BiLSTM network to transmit the timing information from the reverse LSTM back to the forward LSTM to alleviate the gradient vanishing problem, improve its convergence speed and performance, and introduce an improved whale algorithm to further optimize the neural network structure, so as to enhance the accuracy of the model. A mass of comparative simulation experiments are carried out in the IEEE14 node and IEEE118 node test systems, and the simulation outcomes indicate that in comparison with the RNN, LSTM, BiLSTM, and WOABiLSTM models, the proposed model has better measurement accuracy, recall, $\mathbf{F 1}$ score, and real-time detection performance.